Supporting START TRANSACTION WITH CONSISTENT [ROCKSDB] SNAPSHOT

Summary:
This adds two features in RocksDB.
1. Supporting START TRANSACTION WITH CONSISTENT SNAPSHOT
2. Getting current binlog position in addition to #1.
With these features, mysqldump can take consistent logical backup.

The second feature is done by START TRANSACTION WITH
CONSISTENT ROCKSDB SNAPSHOT. This is Facebook's extension, and
it works like existing START TRANSACTION WITH CONSISTENT INNODB SNAPSHOT.

This diff changed some existing codebase/behaviors.

- Original Facebook-MySQL always started InnoDB transaction
regardless of engine clause. For example, START TRANSACTION WITH
CONSISTENT MYISAM SNAPSHOT was accepted but it actually started
InnoDB transaction, not MyISAM. This patch does not allow
setting engine that does not support consistent snapshot.

mysql> start transaction with consistent myisam snapshot;
ERROR 1105 (HY000): Consistent Snapshot is not supported for this engine

Currently only InnoDB and RocksDB support consistent snapshot.
To check engines, I modified sql/sql_yacc.yy, trans_begin()
and ha_start_consistent_snapshot() to pass handlerton.

- Changed constant name from
  MYSQL_START_TRANS_OPT_WITH_CONS_INNODB_SNAPSHOT to
MYSQL_START_TRANS_OPT_WITH_CONS_ENGINE_SNAPSHOT, because it's no longer
InnoDB dependent.

- When not setting engine, START TRANSACTION WITH CONSISTENT SNAPSHOT
takes both InnoDB and RocksDB snapshots, and both InnoDB and RocksDB
participate in transaction. When executing COMMIT, both InnoDB and
RocksDB modifications are committed. Remember that XA is not supported yet,
so mixing engines is not recommended anyway.

- When setting engine, START TRANSACTION WITH CONSISTENT.. takes
snapshot for the specified engine only. But it starts both
InnoDB and RocksDB transactions.

Test Plan: mtr --suite=rocksdb,rocksdb_rpl, --repeat=3

Reviewers: hermanlee4, jonahcohen, jtolmer, tian.xia, maykov, spetrunia

Reviewed By: spetrunia

Subscribers: steaphan

Differential Revision: https://reviews.facebook.net/D32355

mysql-test/suite/rocksdb/r/cons_snapshot_repeatable_read.result

@@ -16,3 +16,110 @@ connection default;
 disconnect con1;
 disconnect con2;
 DROP TABLE t1;
+connect  con1,localhost,root,,;
+connect  con2,localhost,root,,;
+connection con1;
+CREATE TABLE r1 (id int primary key, value int, value2 int) engine=ROCKSDB;
+insert into r1 values (1,1,1),(2,2,2),(3,3,3),(4,4,4);
+BEGIN;
+connection con2;
+INSERT INTO r1 values (5,5,5);
+connection con1;
+SELECT * FROM r1;
+id	value	value2
+1	1	1
+2	2	2
+3	3	3
+4	4	4
+5	5	5
+connection con2;
+INSERT INTO r1 values (6,6,6);
+connection con1;
+SELECT * FROM r1;
+id	value	value2
+1	1	1
+2	2	2
+3	3	3
+4	4	4
+5	5	5
+COMMIT;
+SELECT * FROM r1;
+id	value	value2
+1	1	1
+2	2	2
+3	3	3
+4	4	4
+5	5	5
+6	6	6
+START TRANSACTION WITH CONSISTENT SNAPSHOT;
+connection con2;
+INSERT INTO r1 values (7,7,7);
+connection con1;
+SELECT * FROM r1;
+id	value	value2
+1	1	1
+2	2	2
+3	3	3
+4	4	4
+5	5	5
+6	6	6
+connection con2;
+INSERT INTO r1 values (8,8,8);
+connection con1;
+SELECT * FROM r1;
+id	value	value2
+1	1	1
+2	2	2
+3	3	3
+4	4	4
+5	5	5
+6	6	6
+COMMIT;
+SELECT * FROM r1;
+id	value	value2
+1	1	1
+2	2	2
+3	3	3
+4	4	4
+5	5	5
+6	6	6
+7	7	7
+8	8	8
+START TRANSACTION WITH CONSISTENT SNAPSHOT;
+connection con2;
+INSERT INTO r1 values (9,9,9);
+connection con1;
+START TRANSACTION WITH CONSISTENT SNAPSHOT;
+connection con2;
+INSERT INTO r1 values (10,10,10);
+connection con1;
+SELECT * FROM r1;
+id	value	value2
+1	1	1
+2	2	2
+3	3	3
+4	4	4
+5	5	5
+6	6	6
+7	7	7
+8	8	8
+9	9	9
+START TRANSACTION WITH CONSISTENT SNAPSHOT;
+INSERT INTO r1 values (11,11,11);
+SELECT * FROM r1;
+id	value	value2
+1	1	1
+2	2	2
+3	3	3
+4	4	4
+5	5	5
+6	6	6
+7	7	7
+8	8	8
+9	9	9
+10	10	10
+11	11	11
+drop table r1;
+connection default;
+disconnect con1;
+disconnect con2;

mysql-test/suite/rocksdb/t/consistent_snapshot.inc

@@ -16,7 +16,7 @@ connect (con2,localhost,root,,);
 
 connection con1;
 
-CREATE TABLE t1 (a INT, pk INT AUTO_INCREMENT PRIMARY KEY) ENGINE=rocksdb;
+CREATE TABLE t1 (a INT, pk INT AUTO_INCREMENT PRIMARY KEY) ENGINE=ROCKSDB;
 eval SET SESSION TRANSACTION ISOLATION LEVEL $trx_isolation;
 
 # While a consistent snapshot transaction is executed,
@@ -38,5 +38,70 @@ disconnect con1;
 disconnect con2;
 DROP TABLE t1;
 
+connect (con1,localhost,root,,);
+connect (con2,localhost,root,,);
+
+connection con1;
+CREATE TABLE r1 (id int primary key, value int, value2 int) engine=ROCKSDB;
+insert into r1 values (1,1,1),(2,2,2),(3,3,3),(4,4,4);
+
+BEGIN;
+
+connection con2;
+INSERT INTO r1 values (5,5,5);
+
+connection con1;
+SELECT * FROM r1; # 5
+
+connection con2;
+INSERT INTO r1 values (6,6,6);
+
+connection con1;
+SELECT * FROM r1; # 5
+COMMIT;
+SELECT * FROM r1; # 6
+
+START TRANSACTION WITH CONSISTENT SNAPSHOT;
+
+connection con2;
+INSERT INTO r1 values (7,7,7);
+
+connection con1;
+SELECT * FROM r1; # 6
+
+connection con2;
+INSERT INTO r1 values (8,8,8);
+
+connection con1;
+SELECT * FROM r1; # 6
+COMMIT;
+SELECT * FROM r1; # 8
+
+START TRANSACTION WITH CONSISTENT SNAPSHOT;
+
+connection con2;
+INSERT INTO r1 values (9,9,9);
+
+connection con1;
+START TRANSACTION WITH CONSISTENT SNAPSHOT;
+
+connection con2;
+INSERT INTO r1 values (10,10,10);
+
+connection con1;
+SELECT * FROM r1; # 9
+
+START TRANSACTION WITH CONSISTENT SNAPSHOT;
+INSERT INTO r1 values (11,11,11);
+SELECT * FROM r1; # self changes should be visible
+
+
+drop table r1;
+
+connection default;
+disconnect con1;
+disconnect con2;
+
+
 --source include/wait_until_count_sessions.inc
 

mysql-test/suite/rocksdb/t/disabled.def

@@ -1,7 +1,6 @@
 type_blob_indexes : MDEV-4097 (Indexes on text/blob fields are not allowed)
 type_text_indexes : MDEV-4097 (Indexes on text/blob fields are not allowed)
-cons_snapshot_repeatable_read : Consistent read does not work
-cons_snapshot_serializable : Consistent read does not work
+cons_snapshot_serializable : Consistent read does not work on serializable
 select_for_update : Not supported?
 select_lock_in_share_mode : Not supported?
 level_read_committed : Not supported?